Inspecting a scaffold after high winds or bad weather

Scheduled inspections happen because they are scheduled. Inspections prompted by an event — high winds, an impact, a substantial alteration — depend on somebody deciding that the event was significant enough, and that decision gets made inconsistently because it is made under no particular prompt.

The threshold question, decided in advance

The recurring difficulty is that "high winds" means whatever the person on site thinks it means. One firm treats any night of notable weather as a trigger; another waits for visible damage. Neither has necessarily got it wrong, but a firm with no stated threshold gets a different answer depending on who is asked, which is the thing worth fixing.

Agreeing a threshold in advance — and a route for site to raise anything below it that looked significant anyway — converts a judgement call made under pressure into a routine.

Why these inspections usually find nothing

Most weather-triggered inspections find nothing, which is precisely why the discipline erodes. After several uneventful ones, the next gets deferred, then skipped. The value is not in the average case; it is that this is the category of inspection most likely to find something serious when it does find something.

Covering every structure, not the obvious one

After an event the instinct is to check the structure somebody is worried about. Where a firm has several standing across different sites, a register of what exists is what makes it possible to say the whole exposure has been checked rather than the memorable part of it.

Common mistakes

• –No agreed threshold, so the trigger depends on who is asked
• –Checking the structure somebody is worried about rather than all of them
• –Discipline eroding because previous weather inspections found nothing
• –No record that a weather inspection was carried out, only that nothing was found
• –Not treating impacts from plant or vehicles as a trigger in the same way
• –No route for site to escalate something that looked significant below the threshold

In practice: a threshold that was actually written down

A firm set its trigger at a stated wind speed for the nearest station, with an instruction that any structure over a certain height or with sheeting attached was checked regardless of whether the threshold was reached. It was not sophisticated and it removed the argument.

The valuable part turned out to be the second clause. Sheeted structures behave differently in wind and were the ones the previous informal approach had been least consistent about, because they did not look damaged from the ground. The written threshold caught a category that judgement had been missing.

Recording that nothing was found

Weather inspections that find nothing frequently generate no record at all, because there is nothing to report. That is the wrong instinct: the record that a structure was checked after a named event is precisely the document that matters if a failure occurs later.

A short entry naming the event, the structures checked and the outcome is enough. What it prevents is being unable to show, afterwards, whether anybody looked.

Worth knowing: the structures most at risk are not the tallest

Attention after weather naturally goes to the biggest structures. Sheeted or netted scaffolds, structures with temporary roofs, and anything with loading attached generally behave less predictably in wind than an unclad structure of the same height — and they are easy to overlook because they may be smaller.

• –Prioritise sheeted, netted and loaded structures regardless of height
• –Check anything recently altered before anything long-standing
• –Record what you decided to check and why, not just what you found

Where this connects: recording that nothing was found

Weather inspections that find nothing frequently generate no record, because there is nothing to report. That instinct is exactly backwards: the document proving a structure was checked after a named event is precisely the one that matters if something fails later, and its absence is indistinguishable from nobody having looked.

A short entry naming the event, the structures checked and the outcome is enough. It costs a few minutes and it is the difference between being able to show a considered response to weather and relying on somebody's recollection that they probably had a look.

• –Record the event, the structures checked and the outcome, even when clear
• –Prioritise clad and recently altered structures whatever their height
• –Agree the trigger threshold in advance so it does not depend on who is asked
• –Treat impacts from plant or vehicles as a trigger in the same way
• –Give site a route to escalate something below the threshold
• –Use the register to confirm every standing structure was covered